• Proceedings of the Korea Multimedia Society Conference
• /
• 2002.11b
• /
• pp.339-342
• /
• 2002

동화상 압축에서 사용되는 움직임 벡터는 영상을 일정한 크기의 블록으로 나누어 이전 영상과 가장 예측 오류가 적은 곳을 지정하여 예측한다. 블록의 이동 정도는 수직. 수평의 선형적인 움직임(translational displacements)을 가정하여 사용하기 때문에 실제 화상에서 자주 나타나는 물체의 확대 또는 축소에 의한 크기 변화. 회전. 일그러짐 등의 변화에 올바르게 예측하지 못하는 문제를 가지고 있다. 본 논문은 블록의 선형적인 이동은 물론이고 4개 노드를 자유롭게 움직임일 수 있는 왜곡된 블록 정합 방식의 움직임 예측 기법에 대하여 소개하고 왜곡 블록 정합 방식에서 나타날 수 있는 기하학적인 오류를 수정하는 보정 방법에 관해 논의한다.

• Proceedings of the Korean Institute of Navigation and Port Research Conference
• /
• 2018.11a
• /
• pp.19-21
• /
• 2018

선원의 행동오류는 해양사고를 야기하는 하나의 직접적인 원인이기 때문에 이를 이해하는 것은 해양사고 예방에 근본이 된다. 선원의 행동오류를 이해하기 위해서는 행동오류를 추정하고 예측할 수 있어야 한다. 본 연구에서는 은닉 마르코브 모델(Hidden Markov Model, HMM)을 이용하여 선원들의 행동오류를 추정하고 예측하였다. 아울러 5가지 선박의 종류 각각에 나타나는 선원들의 행동오류를 서로 비교 분석하였다. 모델에 사용한 데이터는 해양안전심판원의 해양사고 보고서에 기록된 내용을 SRKBB(Skill-, Rule- and Knowledge-Based Behavior) 모델을 기반으로 분류하고 관측 수열을 생성하며 라벨링 작업을 통해서 구축하였다. 구축한 데이터를 적용하여 HMM을 보정하고 파라미터를 획득하여 선원들의 행동오류에 관한 모델을 구축하였다. 실험 결과, 선박 종류별로 선원들의 행동오류의 패턴은 서로 다르고, 이를 통해서 선박종류별 해기사들의 행동오류의 추정과 예측이 가능함을 일차적으로 확인할 수 있었다. 추후 본 연구를 지속 전개하여 해양사고 예방을 위한 인적오류의 저감에 기여할 수 있는 방안을 모색할 에정이다.

• The KIPS Transactions:PartB
• /
• v.17B no.4
• /
• pp.275-286
• /
• 2010

Reversible watermarking which can preserve the original quality of the digital contents and protect the copyright has been studied actively. Especially, in medical, military, and art fields, the need for reversible watermarking is increasing. This paper proposes a high-capacity reversible watermarking through predicted error expansion and error estimation compensation. Watermark is embedded by expanding the difference histogram between the original value and the predicted value. Differently from previous methods calculating the difference between adjacent pixels, the presented method calculates the difference between the original value and the predicted value, and that increases the number of the histogram value, where the watermark is embedded. As a result, the high capacity is achieved. The inserted watermark is extracted by restoring the histogram between the original value and the predicted value. To prove the performance, the presented algorithm is compared with other previous methods on various test images. The result supports that the presented algorithm has a perfect reversibility, a high image quality, and a high capacity.

• Proceedings of the Korea Water Resources Association Conference
• /
• 2018.05a
• /
• pp.41-41
• /
• 2018

하천과 저수지의 수질을 예측하고 관리하는데 수리 수질예측모형이 널리 활용되고 있다. 수질예측모형은 유역이나 수체 내의 오염물질 이동경로나 농도를 수치해석 방법으로 계산하여 사용자가 필요로 하는 지점과 시점에서의 수질자료 생산하는데 활용되고 있다. 수질예측모형은 검 보정을 통해 정확도를 확보하며, 정확도의 확보를 위해서는 높은 수준의 전문성을 필요로 한다. 특히 시행착오법으로 모형을 보정하는 경우 많은 시간과 노력을 필요로 하게 되며, 보정계수를 과대 혹은 과소로 모형에 적용하는 오류를 범하기 쉽고 모델러의 주관이 관여되기 쉽다. 그래서 본 연구에서는 CE-QUAL-W2모형의 조류항목에 대한 모형 보정을 위하여 Chl-a와 남조류세포수에서 주로 활용되고 있는 보정계수에 대한 민감도 분석 결과를 토대로 매개변수별 모의결과 변화율을 산정하였으며, 시기적 경향성을 재현하기 위해 Ensemble-Bagging 기법과 머신 러닝 기법을 적용하여 모형 구동횟수를 최소화 할 수 있는 방법으로 구성하였다. Chl-a를 보정하기 위한 매개변수는 9개를 선정하였으며, 규조류, 남조류, 녹조류에 총 27개 매개 변수를 민감도 분석으로 도출 한 후 예상 변화율 대비 이벤트별 모의치와 실측치 간 %difference가 유사하도록 매개변수를 조정하였다. 또한 각 이벤트 조합의 매개변수 빈도수와 매개변수별 예상변화율, 시기적 조류특성을 고려하여 가중치를 도출하였으며, 1회 보정에 맞춰 Chl-a 모델 실행결과를 %difference로 평가한 후 "good"등급을 만족할 때까지 반복 적용하였다. 남조류세포수의 경우 Chl-a에 맞춰 매개변수 최적화 이후 남조류세포수 농도를 세포수로 환산하기 위한 CACEL에 대해 머신러닝 기법을 적용하였으며, CACEL 추정변화율 회귀식에 따라 평가 한 후 %difference "good"등급 이상을 만족할 때까지 반복 수행하는 방법을 적용하였다. 본 연구에서는 수질예측모형의 정확도를 확보하기 위하여 최적화 기법을 적용하였으며, 이를 통해 모형을 보정하는 과정에서 요구되는 시간과 노력을 줄일 수 있도록 하였으며, Ensemble기법과 머신러닝 기법을 적용하여 모형보정계수 적용에 객관성을 확보할 수 있도록 하였다.

• Journal of KIISE:Software and Applications
• /
• v.37 no.6
• /
• pp.417-429
• /
• 2010

Reversible watermarking inserts watermark into digital media in such a way that visual transparency is preserved and then enables to restore the original media from the marked one without any loss of media quality. This watermarking can be applied to quality-sensitive imaging such as medical imaging, military imaging, remote-sensing imaging, and precious artwork, where the original media should be preserved during image processing and analysis. In this paper, a reversible image watermarking technique that embeds message bits by modifying the differential histogram of adjacent pixels is presented. In order to satisfy both high embedding capacity and visual quality, the proposed technique exploits the fact that adjacent pixels in the image have highly spatial correlation. Also, we prevent overflow/underflow problem and salt-and-pepper artifacts by employing a predicted error compensation scheme. Through experiments using various test images, we prove that the presented technique provides perfect reversibility and high embedding capacity, while maintaining the induced-distortion low.

• Journal of the Korea Institute of Information and Communication Engineering
• /
• v.21 no.2
• /
• pp.259-264
• /
• 2017

Birth weight of a new born baby provides very important information in evaluating many clinical issues such as fetal growth restriction. This paper analyzes birth weight data of babies born in Korea from 2011 to 2013, and it shows that there is a biologically implausible distribution of birth weights in the data. This implies that some errors may be generated in the data collection process. In particular, this paper analyzes the relationship between gestational period and birth weight, and it is shown that the birth weight data mostly of gestational periods from 28 to 32 weeks have noticeable errors. Therefore, this paper employs the finite Gaussian mixture model to classify the collected data points into two classes: non-corrupted and corrupted. After the classification the paper removes data points that have been predicted to be corrupted. This adjustment scheme provides more natural and medically plausible percentile values of birth weights for all the gestational periods.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.29 no.1C
• /
• pp.72-82
• /
• 2004

This paper proposes more robust error concealment techniques (ECTs) for MPEG-2 intra coded frame. MPEG-2 source coding algorithm is very sensitive to transmission errors due to the use of variable-length coding. The transmission errors are corrected by error correction scheme, however, they cannot be revised properly. Error concealment (EC) is used to conceal the errors which are not corrected by error correction and to provide minimum visual distortion at the decoder. If errors are generated in intra coded frame, that is the starting frame of GOP, they are propagated to other inter coded frames due to the nature of motion compensated prediction coding. Such propagation of error may cause severe visual distortion. The proposed algorithm in this paper utilizes the spatio-temporal information of neighboring inter coded frames to conceal the successive slices errors occurred in I-frame. The proposed method also overcomes the problems that previous ECTs reside. The proposed algorithm generates consistent performance even in network where the violent transmission errors frequently occur. Algorithm is performed in MPEG-2 video codec and we can confirm that the proposed algorithm provides less visible distortion and higher PSNR than other approaches through simulations.

• Aerospace Engineering and Technology
• /
• v.13 no.1
• /
• pp.27-36
• /
• 2014

Generally, satellites continuously monitor that its major functions are working properly and their hardware are in a good status using several SOH data. In case a fault that is not recognized as a temporal problem or a failure that can be considered to propagate its damage to the other parts are detected, fault management logic is performed automatically without any contact of ground station. In this paper, attitude error detection using sun sensors on a rotating solar array is proposed. Attitude error can be detected by comparing the offset angle between the actual data computed from the sun sensor and the data predicted from the orbit and ephemeris information for the two types of solar array operation method. During the eclipse, the output of attitude error detection method becomes zero because the sun sensor output cannot be provided. Finally, the proposed method is analyzed through the data processing using on-orbit data.

• Journal of the Korea Computer Graphics Society
• /
• v.23 no.2
• /
• pp.41-47
• /
• 2017

In this paper, we present a method for data-driven correction of the noisy motion data captured from a low-end RGB-D camera such as the Kinect device. For this purpose, our key idea is to construct a synchronized motion database captured with Kinect and additional specialized motion capture device simultaneously, so that the database contains a set of erroneous poses from Kinect and their corresponding correct poses from the mocap device together. In runtime, given motion captured data from Kinect, we search the similar K candidate Kinect poses from the database, and synthesize a new motion only by using their corresponding poses from the mocap device. We present how to build such motion database effectively, and provide a method for querying and searching a desired motion from the database. We also adapt the laze learning framework to synthesize the corrected poses from the querying results.

• The Journal of Korean Institute of Communications and Information Sciences
• /
• v.24 no.8B
• /
• pp.1480-1489
• /
• 1999

In this paper, we propose a new post-processing method which can solve a problem of MSDS(Mean Squared Difference of Slope) method. Using that method the blocking artifacts can significantly be reduced without any restriction, which is a major drawback of block-based DCT compression method. In this approach, the OSLD(Overlapped Sub-Laplacian Distribution) of dequantized block boundary pixel difference values is defined and used to categorize each block of an image into one of four types. Those types are also classified into one of two classes: an edge and a non-edge classes. A slope across the block boundary is used to quantify discontinuity of the image. If an absolute estimated quantization error value of a DCT coefficient is greater than the corresponding quantization step size, it is saturated to the step size in the edge class. The proposed post-processing method can improve not only the PSNR value up to 0.1~O.3 dB but visual quality without any constraints determined by ad-hoc manner.